Howtheworld `s school systems - Cowen Institute for Public

Howtheworld `s school systems - Cowen Institute for Public

g
n
i
s
d
m
’
l
r
r
o
w
s
o
m
e
e
f
h
t
s
t
y
r
How -psechool s
p
o
t
t
n
o
s
t
e
u
o
b come
7
0
0
2
r
e
b
m
e
t
p
se
Executive
Summary
Educa­tion reform is top of the agenda of almost every country in the world. Yet despite massive increases in spending (last year, the
world’s governments spent $2 trillion on education) and ambitious attempts at reform, the performance of many school systems has
barely improved in decades. This is all the more surprising because there are wide variations in the quality of education.
For instance, in international assessments, less than one percent of African and Middle Eastern children perform at or above the
Singaporean average. Nor is this solely the result of the level of investment. Singapore, one of the world’s top performers, spends less
on primary education than do 27 of the 30 countries in the OECD.1
Changing what happens in the hearts an minds of millions of children – the main charge of any school system – is no simple task.
That some do so successfully while others do not is indisputable. So why is it that some school systems consistently perform better
and improve faster than others?
There are many different ways to improve a school system, and the complexity of this task and the uncertainty about outcomes
is rightly reflected in the international debate about how this should best be done. To find out why some schools succeed where
others do not, we studied twenty-five of the world’s school systems, including ten of the top performers. We examined what these
high-performing school systems have in common and what tools they use to improve student outcomes.
The experiences of these top school systems suggests that three things matter most: 1) getting the right people to
become teachers, 2) developing them into effective instructors and, 3) ensuring that the system is able to deliver
the best possible instruction for every child.
These systems demonstrate that the best practices for achieving these three things work irrespective of the culture in which
they are applied. They demonstrate that substantial improvement in outcomes is possible in a short period of time and that
applying these best practices universally could have enormous impact in improving failing school systems, wherever they
might be located.
Contents
Foreword
Preface
Introduction: Inside the black box
1.“The quality of an education system cannot
exceed the quality of its teachers”
2.“The only way to improve outcomes is to
improve instruction”
3. “High performance requires every child
to succeed”
Conclusion: The system and the journey
Bibliography
References
Acknowledgements
The authors deeply thank the following educationalists for their counsel and thought partnership:
Michael Fullan, Andreas Schleicher, Lee Sing Kong, S. Gopinathan, and Peter Hill. The authors
would also like to acknowledge the substantial contribution of Fenton Whelan to this report,
the insightful input from McKinsey colleagues Andrew Moffit, Maisie O’Flanagan, Paul Jansen,
the editorial skills of Ivan Hutnik and art director Nicholas Dehaney Media & Design, London.
Foreword
The capacity of countries - both the world’s most advanced economies as well those experiencing rapid development - to compete in the global knowledge economy
increasingly depends on whether they can meet a fast-growing demand for high-level skills. This, in turn, hinges on significant improvements in the quality of schooling
outcomes and a more equitable distribution in learning opportunities.
International comparisons, such as the OECD’s Programme for International Student Assessment (PISA) make it now possible to regularly and directly compare the quality of
educational outcomes across education systems. They reveal wide differences in the extent to which countries succeed in fostering knowledge and skills in key subject areas.
For some countries, results from PISA have been disappointing, showing that their 15-year-olds’ performance lags considerably behind that of other countries, sometimes by
the equivalent of several years of schooling and sometimes despite high investments in education. International comparisons have also highlighted significant variation in the
performance of schools and raised strong concerns about equity in the distribution of learning opportunities. Last but not least, they suggest that there is significant scope for
improving educational efficiency such that, across OECD countries, taxpayers could expect 22% more output for their current investments into schooling.
However, comparisons like PISA also provide very encouraging insights. Across the globe - whether it is Canada in North America, Finland in Europe or Japan and Korea
in Asia - some education systems demonstrate that excellence in education is an attainable goal, and at reasonable cost. They also show that the challenge of achieving a high
and socially equitable distribution of learning outcomes can be successfully addressed and that excellence can be achieved consistently throughout the education systems,
with very few students and schools left behind.
But measuring performance does not automatically lead to insights as to what policy and practice can do to help students to learn better, teachers to teach better, and schools
to operate more effectively. This is where McKinsey’s report comes in, with its first-of-its-kind approach that links quantitative results with qualitative insights on what
high-performing and rapidly improving school systems have in common. With a focus on issues that transcends cultural and socio-economic contexts, such as getting
the right people to become teachers, developing those people into effective instructors, and putting in place targeted support to ensure that every child can benefit from
high-quality instruction, the report allows policy-makers to learn about features of successful systems without copying systems in their entirety.
By enabling policy-makers to examine their own education systems in the light the best performing systems that set the standards of what can be achieved, the report
provides policy-makers with a unique tool to bring about improvements in schooling and better preparation for young people as they enter an adult life of rapid change and
deepening global interdependence. Comparative analyses of this kind will become ever more important, as the best performing education systems, not simply improvement by
national standards, will increasingly become the yardstick for success. Countries will not simply need to match the performance of these countries but do better if their citizens
want to justify higher wages. The world is indifferent to tradition and past reputations, unforgiving of frailty and ignorant of custom or practice. Success will go to those individuals
and countries which are swift to adapt, slow to complain and open to change. The task for governments will be to ensure that countries rise to this challenge.
Andreas Schleicher
Head, Indicators and Analysis Division,
Directorate for Education, OECD
Preface
This report is the result of research carried out by
McKinsey & Company between May 2006 and March 2007.
Its objective has been to understand why the world’s
top-performing school systems perform so very much better
than most others and why some educational reforms succeed
so spectacularly, when most others fail.
Our focus is primarily on how differences in what is
happening at the level of the school system impacts
what is happening in the classrooms, in terms of enabling
better teaching and greater learning. We have chosen not
to focus on pedagogy or curricula, however important
these subjects might be in themselves. These subjects
are well-debated in the literature. There is much less
focus elsewhere on the school ‘system’ itself – the critical
infrastructure that underpins performance – and how to
ensure that it delivers great education for every child.
The report is the outcome of an analysis of the achievements
of the best-performing school systems as defined by the
OECD’s Programme for International Student Assessment
(PISA), a survey of the current literature,2 and interviews
with more than one hundred experts, policymakers and
practitioners. In the course of this research we have visited
schools from Wellington to Helsinki and from Singapore to
Boston in order to benchmark more than two dozen school
systems in Asia, Europe, North America and the Middle East.
The school systems we have benchmarked were selected
to represent two different categories in order to balance the
analysis of current high achievement with developing an
understanding of the route by which others can get there
(Exhibit 1). The first group includes the world’s top ten best-performing school systems
according to the OECD’s Programme for International Student Assessment (PISA); the
second group comprises those that are improving rapidly, having recently introduced
reforms that are raising student outcomes. The examples highlighted throughout this report
are derived from the experiences of these two categories.
We also examined, though to a lesser extent, a third group of school systems located in
developing economies in the Middle East and Latin America that are seeking to provide for
growing populations (Bahrain, Brazil, Qatar, Saudi Arabia, and UAE). This group is currently
embarking on ambitious improvement programs and, in the spirit of focusing on how others
can learn from past experience, we have sought to understand the rationale of their reforms
and how they are adapting approaches that have been successful elsewhere.
Our hope is that this report will help inform the international debate about how to improve
the quality of schools and help chart the path to make future reforms more effective in
improving the quality of schooling for all children everywhere.
9
Introduction:
inside the black
box
Despite substantial increases in spending and many
well-intentioned reform efforts, performance in a large
number of school systems has barely improved in
decades. Few of the most widely supported reform
strategies (for instance, giving schools more autonomy,
or reducing class sizes) have produced the results
promised for them. Yet some school systems
consistently perform better and improve faster than
others. We studied 25 of the world’s school systems,
including 10 of the top performers, to find out why.
Spending, reforms and outcomes
Between 1980 and 2005, public spending per student
increased by 73 percent in the United States of America,
after allowing for inflation. Over the same period, the
U.S. employed more teachers: the student-to-teacher
ratio fell by 18 percent and by 2005, class sizes in the
nation’s public schools were the smallest they had ever
been. The federal government, state governments,
school boards, principals, teachers, teacher unions,
listed companies, non-profit organizations, and others
launched tens of thousands of initiatives aimed at
improving the quality of education in the nation’s schools.
Actual student outcomes, however, as measured by the
Department of Education’s own national assessment
program, stayed almost the same. Though there was
some improvement in mathematics, the reading scores
of 9 year-olds, 13 year-olds and 17 year-olds remained
the same in 2005 as they had been in 1980 (Exhibit 2).
despite substantial
increases in spending and
many well-intentioned
reform efforts, performance
in a large number of
school systems has barely
improved in decades
The United States was not the only country which had
trouble improving its school system. In fact, almost
every country in the OECD substantially increased
its spending on education over the same period, in
addition to launching multiple initiatives to spend this
money more effectively. Yet very few of the school
systems in the OECD achieved significant improvements in performance. One study based on the results
of national and international assessments showed that in
many school systems performance had either flat-lined
or deteriorated (Exhibit 3).3
Yet many of these reform efforts appear well thought-out
and far-reaching in their objectives, making their failure
all the more perplexing. In England, for example, almost
every aspect of the various reforms was reviewed and
reorganized. They reformed “the funding of schools,
the governance of schools, curriculum standards,
assessment and testing, the inspection of quality, the role
of local government, the role of national government,
the range and nature of national agencies, the relationship of schools to communities, school admissions...”4
Yet a report published by the National Foundation for
Education Research in 1996 demonstrated that between
1948 and 1996, despite 50 years of reform, there had
been no measurable improvement in standards of
literacy and numeracy in English primary schools.5
3 Pritchett, Educational Quality and Costs: A Big Puzzle and Five Possible Pieces (2004) | 4 Barber, Journeys of Discovery
(2005) | 5 NFER, Trends in Standards in Literacy and Numeracy in the UnitedKingdom (1997)
it was naive to assume that classroom
quality would improve just because we changed
our structure
The reforms in the United States already mentioned were
similarly ambitious and were concerned with far more
than merely improving the student-teacher ratio.
They also experimented with structural reforms, most
prominently, in the decentralization of powers in school
districts, smaller schools, and charter schools (schools
given increased autonomy in exchange for increased
accountability). Yet the results were disappointing.
Though the best charter schools demonstrated significant improvements in student outcomes were possible,
and certain chains of charter schools showed that
reliable models could consistently deliver improvements
in a succession of schools, in the aggregate, the results
of the charter schools did not significantly outperform
those of other schools. The National Assessment of
Educational Progress (NAEP) went so far as to suggest
that students in charter schools slightly underperformed
their counterparts in public schools, even after
allowing for student background (Exhibit 4).6 Similarly,
‘small schools’ (new schools created by breaking up
larger high schools) showed “slightly improved results
in reading, and worse results in math.”7
11
6 NAEP, America’s Charter Schools: Results from the NAEP Pilot Study (2003)
7 Business Week, Bill Gates Gets Schooled, (2006)
8 Interview: New Zealand, May, 2006
9 Cross City Campaign, A Delicate Balance: District policies and Classroom Practice (2005)
10 OECD, Attracting Developing and Retaining Effective Teachers (2005).
11 Hanushek, The Evidence on Class Size (2003). Shapson, An experimental study on the effects of class size.
Akerhielm, Does class size matter?
12 Ibid
In New Zealand, policymakers overhauled the
structure of the system, decentralizing powers to
individual schools (which would be governed by
elected boards), created two new independent
regulatory bodies, and significantly reduced the role of
central government in the school system. Five years on,
in the mid-1990s, up to one third of schools were failing.
One policymaker explained, “It was naive to assume
that classroom quality would improve just because we
changed our structure.”8
A report by the Cross City Campaign, which analyzed
similar reforms in Chicago, Milwaukee and Seattle,
concluded that, “The three districts had decentralized
resources and authority to the schools in different ways
and had undergone significant organizational changes
to facilitate their ambitious instructional improvement
plans. The unfortunate reality for the many principals
and teachers we interviewed is that the districts were
unable to change and improve practice on a large scale.
And the evidence is indisputable: you can’t improve
student learning without improving instruction.”9
The one policy that almost every school system has
pursued is in reducing class sizes. “Class size reduction,
facilitated by lower student-to-teacher ratios, has
probably been the most widely supported and most
extensively funded policy aimed at improving schools.”10
Over the past five years every country in the OECD
except for one has increased the number of its teachers
relative to the number of its students.
Yet the available evidence suggests that, except at the
very early grades, class size reduction does not have
much impact on student outcomes. Of 112 studies
which looked at the impact of the reduction in class
sizes on student outcomes, only 9 found any positive
relationship. 103 found either no significant relationship,
or a significant negative relationship.11 Even when a
significant relationship was found, the effect was not
substantial. More importantly, every single one of the
studies showed that within the range of class sizes
typical in OECD countries, “variations in teacher
quality completely dominate any effect of reduced
class size”.12 Moreover reducing class sizes had
significant resource implications: smaller classes
ten years ago, seminal research based on
data from Tennessee showed that if two
average eight-year-old students were
given different teachers – one of them a
high performer, the other a low performer
– their performances diverge by more than
50 percentile points within three years
meant that the school systems needed more teachers,
which in turn meant that, with the same level of
funding, they had less money per teacher. It also
meant that because the school system requires more
teachers to achieve smaller class sizes it could become
less selective about who could be a teacher.13
Focusing on teacher quality
The available evidence suggests that the main driver of
the variation in student learning at school is the quality
of the teachers. Ten years ago, seminal research based
on data from Tennessee showed that if two average
eight-year-old students were given different teachers –
one of them a high performer, the other a low performer
– their performance diverge by more than 50 percentile
points within three years (Exhibit 5).14
By way of comparison, the evidence shows that reducing
class sizes from 23 to 15 students improves the performance of an average student by eight percentile points
at best.15 Another study, this time in Dallas, shows that
the performance gap between students assigned three
effective teachers in a row, and those assigned three
ineffective teachers in a row, was 49 percentile points.16
In Boston, students placed with top-performing math
teachers made substantial gains, while students placed
with the worst teachers regressed – their math got
worse.17 Studies that take into account all of the available
evidence on teacher effectiveness suggest that students
placed with high-performing teachers will progress
three times as fast as those placed with low-performing teachers.18 In every school system visited during the
benchmarking, head teachers reported variations in the
amount of learning that occurred in different classes,
and those variations depended mainly on the quality of
teaching in different classrooms. The negative impact
of low-performing teachers is severe, particularly during
the earlier years of schooling. At the primary level,
students that are placed with low-performing
teachers for several years in a row suffer an educational
loss which is largely irreversible. In some systems, by
age seven, children who score in the top 20 percent on
tests of numeracy and literacy are already twice as likely
to complete a university degree as children in the
bottom 20 percent. In England, students that were failing
at age 11 had only a 25 percent chance of meeting the
standard at age 14. By age 14, the chances that a failing
student would graduate with the expected minimum set
of school-leaving qualifications had fallen to just six
percent (Exhibit 6). Taken together, all the evidence
suggests that even in good systems, students that do
not progress quickly during their first years at school,
because they are not exposed to teachers of sufficient
calibre, stand very little chance of recovering the lost
years.
13 The most optimistic estimates of the effectiveness of reducing class size on student achievement suggest that
a reduction in class size from 23 to 15 in the early grades leads to an improvement in performance equivalent to 0.2
standard deviations. | 14 Sanders & Rivers, Cumulative and Residual Effects of Teachers on Future Student Academic
Achievement (1996). | 15 Scientific American, Does Class Size Matter (2001). | 16 Teacher Effects on Student
Achievement (1997) 17 Kati Haycock, Achievement in America: Can we close the gaps (2006)
Striking differences,
fundamental similarities
13
Yet some school systems do perform better and improve
faster than others. Singaporean students score top in
the TIMSS assessment (an international examination in
Mathematics and Science) despite the fact that Singapore
spends less on each student in primary education than
almost any other developed country. In Finland, students
do not start school until they are seven years old, and
attend classes for only four or five hours each day
during their first two years of schooling. Yet by age
15, they score top in the world in tests of mathematics,
science, reading and problem solving, a full 50 points
ahead of their peers in neighbouring Norway. In the
United States, Boston increased the number of students
meeting the MCAS standard from 25 percent to
74 percent in Math, and from 43 percent to 77 percent
in English, in just six years.
Clearly there are inevitable differences between
schools: policy makers in Seoul, Helsinki and Chicago
operate in completely different cultural and political
contexts, and confront different challenges. Some
systems appear to be polar opposites: the Netherlands
attributed much of their success to a highly devolved
governance system; Singapore says it succeeded
because of strong central control; England’s system
contains 23,000 schools, Boston’s just 150.
Yet there were also fundamental similarities. We found
that high-performing school systems, though strikingly
different in construct and context, maintained a strong
focus on improving instruction because of its direct
impact upon student achievement. To improve
instruction, these high-performing school systems
consistently do three things well:
They get the right people to become teachers
(the quality of an education system cannot exceed
the quality of its teachers).
They develop these people into effective
instructors (the only way to improve outcomes
is to improve instruction).
They put in place systems and targeted support
to ensure that every child is able to benefit from
excellent instruction (the only way for the system
to reach the highest performance is to raise the
standard of every student).
Acting on these drivers requires that changes and
improvements be made in other parts of the system,
ranging from funding structures to governance and
incentives. These systems all ensure that they put in
place the necessary foundational conditions, such as
rigorous standards and assessments, clear expectations,
differentiated support for teachers and students, and
sufficient funding, facilities and other core resources.
So, although it is true that the system’s context, culture,
politics and governance will determine the course which
system leaders must follow, the cumulative experience
of the high-performing systems we studied indicates
that focusing on these three drivers is essential for
improving student outcomes and, more importantly, that
reform efforts which fail to address these drivers are
unlikely to deliver the improvements in outcomes that
system leaders are striving to achieve. The remainder
of this report explores these drivers in more detail.
15
T
“
he quality of an education
system cannot exceed the
quality of its teachers”
The top-performing school systems consistently attract
more able people into the teaching profession, leading to
better student outcomes. They do this by making entry
to teacher training highly selective, developing effective
processes for selecting the right applicants to become
teachers, and paying good (but not great) starting
compensation. Getting these essentials right drives up
the status of the profession, enabling it to attract even
better candidates.
The quality of a school system rests on the quality of its
teachers. The evidence that getting the right people to
become teachers is critical to high performance is both
anecdotal and statistical. A South Korean policymaker is
explicit about the importance of getting good people
into teaching: “The quality of an education system
cannot exceed the quality of its teachers”.19 In the United
States, studies show that “a teacher’s level of literacy, as
measured by vocabulary and other standardized tests,
affects student achievement more than any other measurable teacher attribute.”20 While it is a matter of debate,
some studies have found that teachers working for Teach
For America (a program which targets graduates of top
universities) get significantly better outcomes from their
students than do other teachers. This is the case despite
the fact that their teachers have only a short period of
teacher training, work in the toughest schools, and
generally have no prior experience (teacher
effectiveness increases dramatically during the first five
years of teaching).21
The top-performing systems we studied recruit their
teachers from the top third of each cohort graduate from
their school system: the top 5 percent in South Korea,
the top 10 percent in Finland, and the top 30 percent
in Singapore and Hong Kong. In the United States,
programs in rapidly improving systems, such as the
Boston Teacher Residency, the New York Teaching
Fellows, and the Chicago Teaching Fellows do the same
thing, targeting the graduates of top universities.
Conversely, lower-performing school systems rarely
attract the right people into teaching. The New
Commission on the Skills of the American Workforce
observes that, “We are now recruiting our teachers from
the bottom third of high-school students going to college
... it is simply not possible for students to graduate [with
the skills they will need]... unless their teachers have the
knowledge and skills we want out children to have.”22
A Middle Eastern policymaker a region where teachers
have historically been recruited from the lowest third
of high-school graduates is succinct: “faakid ashay la
yua’tee” (“One cannot give what one does not have”).23
19 Interview: South Korea, 2007 | 20 NCTQ, Increasing the Odds: How good policies can yield better teachers | 21 Decker, Mayer, Glazerman,
The Effects of Teach for America: Findings from a National Evaluation (2004) | 22 NCEE, Tough Choices or Tough Times (2007) | 23 Interview: GCC, May 2006 | 24 Training and Development Agency for Schools (11 August 2005)
Culture, policy and the status
of teaching
In all of the systems we studied, both policymakers and
commentators frequently attributed their success in
attracting talented people into teaching (or the lack
thereof) to variables seemingly outside the control of the
policymaker: history, culture, and the status of the
teaching profession. In particular, outsiders often attribute
the success of the Asian school systems we studied to the
dual blessing of a high cultural premium on education
and traditional (Confucian) respect for teachers.
Despite this common belief, our benchmarking
suggests that the same broad policies are effective in
different school systems irrespective of the cultural
context in which they are applied. School systems in
Europe and America which have made the same policy
choices as Asian school systems attract the same quality of applicants, or better: the Chicago Teaching Fellows
and Boston Teacher Residency, for instance, attract the
same calibre of graduate as Singapore or Hong Kong.
Some school systems have made strategic policy interventions that have quickly transformed the status of the
teaching profession: England has made teaching the
most popular profession among undergraduates and
graduates in just five years.24 Even in systems where
the teaching profession enjoys a traditionally high status,
policy still had a massive impact on quality. Finland has
lifted the status of its primary school teachers relative to
17
those in secondary schools by varying salaries by as
little as €100 a month. In South Korea there is a substantial
difference between the status of primary teachers and
secondary teachers: this is entirely attributable to government policy in controlling the supply in teacher training
places for primary school teachers. In each system we
studied the evidence suggests that policies have a strong
impact on status, irrespective of the cultural context in
which they are applied.
Looking at the various systems as a whole, there are
common strategies and best practices for attracting
strong candidates into the teaching profession. England
has led the way in using marketing and recruitment
techniques taken from business to increase the supply
of quality applicants. Most top-performing school
systems remove obstacles to entry into the profession
by creating alternative pathways for experienced hires.
Most of the systems also recognise that they will make
mistakes, and have developed processes to remove
low-performing teachers from the classroom soon after
appointment.
Almost universally, the top school systems do two things:
they have developed effective mechanisms for selecting
teachers for teacher training, and they pay good starting
compensation. These two things have a clear and
demonstrable impact on the quality of people who
become teachers. These same features are frequently
absent in lower-performing systems.
Mechanisms for selecting
teachers for teacher training
The top-performing school systems have more
effective mechanisms for selecting people for teacher
training than do the lower-performing systems.
They recognize that a bad selection decision can result
in up to 40 years of poor teaching. These mechanisms
acknowledge that for a person to become an effective
teacher they need to possess a certain set of characteristics that can be identified before they enter teaching:
a high overall level of literacy and numeracy, strong
interpersonal and communications skills, a willingness
to learn, and the motivation to teach.25 The selection
procedures are therefore designed to test for these
skills and attributes, and select those applicants that
possess them. Singapore’s and Finland’s selection
procedures are among the most effective. Both these
systems place a strong emphasis on the academic
achievement of candidates, their communication skills,
and their motivation for teaching. Singapore has implemented a single, state-wide selection process that is
managed jointly by the Ministry of Education and the
National Institute for Education (Exhibit 7).
Finland has introduced a national first-round in its
selection process which, from 2007 onwards, will
consist of a multiple-choice examination designed to
test numeracy, literacy and problem-solving skills.26
The top-scoring candidates are then passed through to
second round in the selection procedure which is run by
the individual universities. In this round the applicants
are tested for their communication skills, willingness
to learn, academic ability, and motivation for teaching.
Upon graduation from teacher training, the prospective
teachers nevertheless need to pass yet further tests run
by the individual schools to which they apply for
teaching positions (Exhibit 8).
the top-performing
school systems have
more effective
mechanisms for
selecting people
for teacher training
than do the
lower-performing
systems
25 Allington, Johnston, What do we know about effective fourth grade teachers and their classrooms (2000). Interviews
in Singapore, South Korea, and Hong Kong | 26 Before 2007, the first round of the recruitment process had been based
mainly on achievement at secondary school.
As the quality of people on the courses begins to drop,
so does the quality of the courses themselves, because
the quality of any classroom experience is highly dependent on the quality of the people in the classroom.
The programs also suffer from having too many students: if the program had selected just the number of
people needed to fill the vacant teaching posts, they
would have been able to spend almost three times as
much on training each student. All told, Option 2 tends
to make teacher training a low-status program, which in
turn makes teaching a low-status profession. Once this
has been allowed to happen, teaching becomes stuck in
a downward spiral.
As important as it is to get the selection process right,
it is equally important to make sure that the selection
process happens at the right point in time. In every
system we studied, teachers begin their professional
careers with a period of teacher training. In most cases
this consisted of either a three- or four-year undergraduate program, or a one-year postgraduate program
following an undergraduate degree in a subject other
than education. School systems therefore have two
options for selecting teachers (Exhibit 9).
Option 1:The first model selects people before
they start their teacher training and limits
places in the training program to those
who are selected.
Conversely, the top-performing systems select for
entry into the teacher training programs. They do so
either by controlling entry directly, or by limiting the
number of places on teacher training courses, so that
supply matches demand. In Singapore, applicants are
screened, tested and selected before they enter teacher
training (Exhibit 10). They are then formally employed
by the Ministry of Education and paid a salary during
their training.27 This means that teacher training is not an
option for those with few other options. Making teacher
training selective in this manner makes it attractive to
high performers. It also means that Singapore can, and
does, spend more on teacher training (per student) than
other education systems because there are fewer people
in its courses. All of this makes teacher training an attractive and high-status course in Singapore and this, in turn,
makes teaching an attractive and high-status profession.
Option 2: The second model leaves the selection
process until after the prospective
teachers have graduated from teacher
training and then selects the best
graduates to become teachers.
While almost every school system in the world uses
the second option, most of the top-performers use
variations on the first.
Failing to control entry into teacher training almost
invariably leads to an oversupply of candidates which,
in turn, has a significant negative effect on teacher
quality. In one system we benchmarked, of 100 people
that applied to teacher training, only 20 became
teachers. Of this 100, 75 received offers for teacher
training places, indicating that it is relatively easy to
get into the teacher training program. However, upon
graduation, because of over-supply, they struggle to
find jobs as teachers, making the course less appealing
to the more able students. In such conditions teacher
training became an option for students who had few
other options available to them.
27 A full salary is paid during training on one-year programs. On longer programs,
a salary is only paid during the final part of the course
Several other school systems have created similar
structures to those seen in Singapore. Finland limits the
number of places on teacher training so that the supply
of teachers matches demand, and only allows universities to select candidates who have passed a national
screening process. Boston, Chicago and New York have
a somewhat different approach in that they control entry
into teacher training only for prospective teachers on
their Fellows and Residency programs (rather than
for all prospective teachers). For these programs
candidates are selected through a system-wide admissions process and guaranteed a teaching position in a
school before they enter teacher training. Both programs
report that the calibre of their candidates is much higher
than the cities’ average intake. England focuses on
limiting the funding for teacher training to manage
supply, and ensures that all training providers meet
certain general standards for the selection of the
students in their courses.
19
A compelling example of how the control of entry to
teacher training programs can have a substantial
positive impact on the quality of people who become
teachers is seen in the contrast between how South
Korea’s system treats its primary school and secondary
school teachers.
In order to become a primary teacher it is necessary
for the prospective teacher to first complete a four-year
undergraduate degree in education at a National
Education University. Places on these courses are
limited, to ensure that the supply of teachers meets
demand. Entry is by merit. Admission to all first degree
courses in South Korea is based on the results of the
national College Entrance Exam; the cut-off score for
teacher training courses requires that students should
be in the top five percent of their academic cohort.
The courses are therefore highly selective and the
graduates of these courses are very likely to find
employment as a teacher. This ensures that the attractiveness, status and quality of the courses remain high.
South Korea takes a very different approach to training
its secondary school teachers, however, resulting in very
different outcomes. In contrast to its careful matching of
supply with demand for primary school teachers, the
selection of secondary school teachers are not subjected
to the same approach. Instead of facing restrictions in
entry to training courses, they are free to complete their
teacher training at one of more than 350 competing providers. Graduates then apply for jobs at one of the 16
provincial or metropolitan offices of education.
As a result, there is significant oversupply: South Korea
produces five times as many graduates each year as is
required by the secondary school system. This problem
has been compounded over time and the number of
applicants now exceeds the number of places by a factor
of eleven (in December 2005 there were 59,090
applications for 5,245 teaching positions). As a result,
in contrast to situation for primary school teaching, the
status and attractiveness of secondary school teaching
has declined in South Korea, making it unattractive to
high-performers.
Selective entry has clear benefits. Broadly, there are
three different mechanisms that school systems use to
make entry into teacher training more selective and to
match the supply of teacher training with demand.
System-wide recruitment processes: In Singapore
and Finland, to different degrees, the state controls
the entire process for the selection of students for
teacher training. In Singapore, prospective teachers
are selected and employed by the Ministry of
Education before entering teacher training. In
Finland, there is a two-stage process. In the first
stage, prospective teachers are subjected to a
nation-wide screening process. In the second stage,
the individual universities select their own candidates
from those that have met the criteria in the first stage.
Places in teacher training courses in both countries
are limited so that the supply of graduates matches
demand.
Controlling places through funding: In Hong Kong,
England and South Korea’s primary school system,
the government uses its control of funding to limit the
number of students (and the supply of teacher training places). This approach assumes that once supply
is restricted, universities will implement rigorous
selection procedures to ensure that the best applicants are selected. This approach probably functions
best in England, which defines the competencies for
new teachers, has a rigorous quality assurance system, and puts in place penalties for under-performing training providers. This ensures that the training
providers have the right incentives to implement
thorough selection processes.
Alternative pathways: Where the system leaders
can not influence the university selection procedures
or funding, the systems have created alternative
entry paths that enable them to select suitable
candidates before their entry into training. The
Boston Teacher Residency, Chicago Teaching
Fellows, and New York Teaching Fellows programs
all follow this approach, guaranteeing those selected
a teaching position before they enter the training
program. These districts have entered into
agreements with the local schools and universities
to provide training for the candidates they select.
In addition to developing alternative ways of recruiting fresh graduates, top-performing systems have also
found ways to recruit more experienced graduates.
Typically, teacher training requirements create barriers
to recruiting such people. Applicants to teaching who
have already completed their university studies and
started work generally have to undertake a year of training, during which they lose a year’s earnings, as well as
often having to bear the cost of their course in addition.
This makes entry into the profession unattractive to experienced hires, particularly those with families or other
financial commitments. Opening up alternative routes
all of the top-performing systems we benchmarked
(except for one) paid starting salaries that
were at or above the OECD average, relative to
their GDP per capita
into teaching in which entrants are relieved of this
financial burden increases significantly the pool of
potential applicants into the profession. Most systems
have also found that the quality of applicants on these
programs is higher than otherwise (Exhibit 11).
England has probably diversified its recruitment
process the most, having developed more entry points
into teaching than any other system in an attempt to
maximise recruitment. By 2006 there were 32 different
ways to enter the teaching profession in England, though
the expectations of the skills, knowledge, and the behaviours teachers should demonstrate by the time they had
completed their training is the same for each route.
28 Interview: Finland, March, 2007
Most top-performing systems recognise that no
selection process is perfect, and so implement procedures to ensure that the lowest-performing teachers
can, if necessary, be removed from the classroom after
appointment to their teaching position, based on the
evidence of their classroom practice. In the rapidly improving systems of Boston and Chicago, teachers are not
made permanent until they have been teaching for three
or four years, respectively. This allows the district to
remove them from their position if they prove unsuitable.
In England and New Zealand teachers do not gain their
teaching licences until after they have completed one or
two years teaching, respectively, and have gained
satisfactory reviews from their principals. In New
Zealand, the Teachers’ Council makes a second,
follow-up evaluation of 10 percent of all new teachers so
as to ensure the evaluations undertaken by the school
principals meet the right standard.
Good starting compensation
The other essential ingredient for getting the right
people to become teachers is to provide good starting
pay. All of the top-performing systems we benchmarked
(except for one) paid starting salaries that were at or
above the OECD average, relative to their GDP per
capita. What is interesting, however, is that the range of
starting salaries offered by the top performers is very
narrow: most systems pay a starting salary between 95
percent and 99 percent of GDP per capita (across the
OECD as a whole, starting salaries range from 44
percent to 186 percent of GDP per capita) (Exhibit 12).
A good salary is not necessarily the main or only
motivation for teaching, of course. Surveys show that
most people who enter the teaching profession do so for
a range of reasons, the foremost of which is the desire to
help a new generation succeed in a world in which skills
and knowledge are crucial to success. In fact, salary is
rarely stated to be one of the most important reasons
for becoming a teacher, even in the systems where
compensation is good; in the words of one Finnish
teacher, “None of us do this for the money”.28 However,
the surveys also show that unless school systems offer
salaries which are in-line with other graduate starting
salaries, these same people do not enter teaching.
This has important implications for policy. Top-performing systems have found that while raising salaries in line
with other graduate salaries is important, raising them
above the market average for graduates does not lead to
substantial further increases in the quality or quantity of
applicants. In England, where salaries had been slightly
21
teaching.29 This might explain why countries which pay
very high starting salaries (in Europe, Spain, Germany
and Switzerland pay the highest starting salaries relative
to GDP) have not gained improved outcomes as a result.
Only in South Korea, where salaries are exceptionally
high (not only do they start high, but they rise to a
maximum that is two-and-a-half times higher than the
average maximum teacher salary in the OECD)30 do
higher salaries appear to have resulted in an increase
in the quality of people becoming teachers.
Clearly, paying higher starting salaries places a financial
burden on the school system. Broadly, there are three
strategies for balancing the cost of paying higher starting salaries:
Spending more: Boston Public Schools pay the
highest starting salaries in Massachusetts. In order
to do so, it spends more: its annual spending on
primary education per student is equivalent to
26 percent of GDP per capita, significantly above
the OECD average. However, most of the top
performers spent less on their school systems than
the OECD average – they have found other ways to
fund higher starting salaries (Exhibit 13).
below average graduate salaries, increasing teacher
salaries by a small amount (10 percent) resulted in a
substantial rise in applications (30 percent); whereas, in
Switzerland, where salaries were already very high (116
percent of GDP per capita), further increases in salary
had little impact on the number or quality of applicants to
F
rontloading compensation: Finland, the Netherlands, New Zealand, Australia and England, in effect,
frontload their compensation: the starting salaries
are good, but relative to other OECD countries,
subsequent increases in compensation are small.31
In Finland, the difference between the average starting salary and the maximum teacher salary is just 18
percent (Exhibit 14). By paying good starting salaries,
Finland attracts strong performers into the profession.
Teachers who are committed to teaching stay despite
the salary; others who are less committed leave, as
their compensation decreases relative to their peers
in other professions. Systems which frontload compensation succeed because of two factors: first, salary progression is less important in the decision to
become a teacher than starting salary and, secondly,
teacher retention is generally not correlated strongly
to salary progression.
Though restructuring salary scales in order to frontload compensation is likely to prove difficult to achieve
in most school systems, it is not impossible. One of the
top-performers, the Netherlands, has done exactly this.
Between 1990 and 1997, the Netherlands increased its
monthly starting salary for teachers from €1,480 to
€2,006, effectively bringing teachers’ starting salaries
into line with the private sector.32 The Netherlands also
reduced the time it takes to reach the top of the salary
schedule from 26 years to 18 years, with the eventual
aim of reducing it to 15 years. Similarly, Alberta has
been increasing its starting salaries more quickly than
its maximum salary, and has reduced the difference
between the top and bottom of its scale from 81 percent
to 70 percent since 2001. Some of the school systems
use other mechanisms to frontload compensation, such
as paying salaries or bursaries during teacher training
(Boston, England, Chicago, New York, Singapore) or
offering signing bonuses to new teachers (England).
Increasing class size: South Korea and Singapore
employ fewer teachers than other systems; in effect, this ensures that they can spend more money
on each teacher at an equivalent funding level. Both
countries recognise that while class size has relatively
little impact on the quality of student outcomes (see
above), teacher quality does. South Korea’s studentto-teacher ratio is 30:1, compared to an OECD average of 17:1,33 enabling it to in effect double teacher
salaries while maintaining the same overall funding
level as other OECD countries (teacher salaries are
29 OECD, Attracting Developing and Retaining Effective teachers, (from Dolton, Wolter, Denzler) p. 70 | 30 Starting primary teacher salaries in South Korea are 141 percent of GDP per capita, rising to 389 percent of GDP per capita
(compared to OECD averages of 95 percent and 159 percent of GDP per capita respectively) (2003). | 31 The increase in the maximum salary over the starting salary in high-performing systems is as follows: an increase of 18 percent in Finland, 45
percent in New Zealand and the Netherlands, 46 percent in England, and 47 percent in Australia (average across all states and territories), compared to an OECD average of 70 percent. Source: OECD, Education at a Glance 2005 | 32 Attracting Developing and
Retaining Teachers: Country Report for the Netherlands, pp. 36-37 | 33 2003 (OECD, Education at a Glance 2005)
In all of the school systems the status of teaching is
driven mainly by policy, and policies can change
its status very quickly. There are two dominant
approaches for changing the status of the profession:
Separate branding: Boston, Chicago, Teach First and
Teach For America have all created distinct brands
with a separate status associated with them. For
instance, Teach First and Teach For America have
successfully branded themselves as programs distinct from mainstream teaching: “Teach First succeeded in making teaching acceptable among a
group who had perceived it as having low status by
constructing the participants as an elite group.”34
the main budget item in any school system budget,
typically representing 60-80 percent of spending).
Singapore has pursued a similar strategy, but has
also frontloaded compensation. This combination
enables it to spend less on primary education than
almost any OECD country and yet still be able to attract strong candidates into the teaching profession. In
addition, because Singapore and South Korea need
fewer teachers, they are also in a position to be more
selective about who becomes a teacher. This, in turn,
increases the status of teaching, making the profession even more attractive.
The importance of teacher status
In all of the systems we studied, the ability of a school
system to attract the right people into teaching is closely
linked to the status of the profession. In Singapore and
South Korea, opinion polls show that the general public
believe that teachers make a greater contribution to society than any other profession. New teachers in all of the
systems studied consistently reported that the status of
the profession is one of the most important factors in their
decision to become a teacher.
In all school systems there are powerful feedback loops
associated with the status of the teaching profession.
Once teaching became a high-status profession, more
talented people became teachers, lifting the status of
the profession even higher. This is particularly apparent
in Finland and South Korea, where historically strong
teaching forces have given the profession a high status
in the eyes of the general public, enabling them to attract
further high-calibre recruits, thereby perpetuating this
status. Conversely, where the profession has a low status,
it attracts less-talented applicants, pushing the status of
the profession down further and, with it, the calibre of
people it is able to attract. The power of these feedback
loops suggests that seemingly small policy changes can
sometimes have a massive impact on the status of the
teaching profession.
System-wide strategies: Singapore and England
have both implemented carefully constructed marketing strategies, linked to recruitment programs,
which have sought to raise the status of the profession.
In both cases, the systems leveraged best-practices
from business. The marketing was backed by
tangible improvements to starting conditions,
particularly increased salaries.
The Training and Development Agency for Schools
(TDA) in England tracked the response to its marketing
campaigns and, based on the feedback it was getting,
carefully modified its approach (Exhibit 15).
The TDA had been given the task of raising the
quality and quantity of applicants into teaching. To
do this it chose to employ best-practice marketing and
recruiting techniques used in business: it carefully
segmented its target audience, tracked individual
candidates through a sophisticated relationship
management system, scripted key interactions
between its representatives and prospective teachers,
and got feedback through surveys and market
research (Exhibit 16). It also supported two differentiated
34 IPSE, An evaluation of innovative approaches to teacher training on the Teach First Programme (2006)
23
in all of the systems we studied, the ability of a school
system to attract the right people into teaching is closely
linked to the status of the profession
programs to appeal to different segments of the market.
Its Teach First35 program targets top university graduates, while FastTrack is designed to attract and develop
potential school leaders.
In addition to changing how the teaching profession is
perceived externally, most systems have found that the
perception of the teaching profession is linked to the
perceived level of education and training that teachers
are required to undertake to become teachers.
Emphasis on development: Policymakers in
Finland have raised the status of the teaching
profession by requiring that all teachers possess a
master’s degree. Singaporean policymakers have
achieved a similar result by ensuring the academic
rigour of their teacher education courses, as well as
by providing all teachers with the entitlement of 100
hours fully-paid professional development training
each year.
Conclusion
The debate about how to improve the world’s school
systems has all too often been guided by a set of beliefs
that have little basis in fact: namely that it is possible to
make substantial long-term improvement to the school
system without fundamentally raising the quality of
people who enter the teaching profession; important
variables, such as the status of the teaching profession,
are largely outside the control of policymakers; attracting
better people into teaching will always require school
systems to pay ever higher salaries; making teaching
the preferred career choice for large numbers of
top-performers is an unattainable, or at best, distant
goal. The experiences of the high-performing school
systems suggest that all these beliefs fail the test of
critical examination.
School systems, from Seoul to Chicago, from London
to New Zealand, and from Helsinki to Singapore, show
that making teaching the preferred career choice
depends less on high salaries or ‘culture’ than it does
on a small set of simple but critical policy choices:
developing strong processes for selecting and
training teachers, paying good starting compensation,
and carefully managing the status of the teaching
profession. Above all, the top performing systems
demonstrate that the quality of an education system
depends ultimately on the quality of its teachers.
35 Teach First targets graduates of the top universities in the United Kingdom, asking them to spend two years teaching. It then
supports them in getting other jobs in the private sector after they had finished two years teaching. Not only are its teachers
highly successful, but 47 percent of the first cohort decided to stay in teaching after the end of the two-year program.
25
Theonly
way
“to
improve outcomes is
to improve instruction
”
The top-performing school systems recognise that
the only way to improve outcomes is to improve
instruction: learning occurs when students and
teachers interact, and thus to improve learning implies
improving the quality of that interaction. They have
understood which interventions are effective in achieving this – coaching classroom practice, moving teacher
training to the classroom, developing stronger school
leaders, and enabling teachers to learn from each other
– and have found ways to deliver these interventions
throughout their school system.
The quality of the outcomes for any school system is
essentially the sum of the quality of the instruction that
its teachers deliver. “You could define the entire task
of [a school] system in this way: its role is to ensure that
when a teacher enters the classroom he or she has
the materials available, along with the knowledge, the
capability and the ambition to take one more child up to
the standard today than she did yesterday. And again
tomorrow.”36 Ensuring that teachers have that knowledge
and capacity is not easy. Delivering excellent instruction
requires teachers to develop a highly sophisticated set
of skills. Alberta’s standards for effective teaching, for
instance, list more than 30 variables that teachers are
expected to consider when deciding which instructional
techniques to use in any given situation. By age nine,
“the achievement gap within a single class may span
five or more years of schooling.”37 Teachers need to be
able to assess precisely the strengths and weaknesses
of each individual student they teach, select the appropriate instructional methods to help them to learn, and
deliver instruction in an effective and efficient manner.
The first part of the challenge is to define what great
instruction looks like. That task – developing the
curriculum and its associated pedagogies – is difficult
and controversial from an educational perspective, yet
relatively more straightforward from a system management perspective: the challenge is broadly one of finding the best educators and giving them the space to
debate and create a better curriculum and pedagogy.
The second part of the challenge in instruction is, at
least from a system management perspective, much
more complex: giving thousands of teachers (in some
cases hundreds of thousands of teachers) the capacity
and knowledge to deliver that great instruction reliably,
every day, across thousands of schools, in circumstances that vary enormously from one classroom to the next
– and all this with very little oversight.
All of the rapidly improving systems recognise the
complexity and primacy of this second challenge, and
focus much of their reform effort on developing and
implementing successful strategies to improve
classroom instruction. One policymaker in Boston
explained that, “The three pillars of the reform were
professional development, professional development,
and professional development... We aligned everything
– resources, organization, people – with professional
you could define the
entire task of (a
school) system in this
way: its role is to
ensure that when a
teacher enters the
classroom he or she has
the materials available,
along with the
knowledge, the
capability and the
ambition to take one
more child up to the
standard today than she
did yesterday.
And again
tomorrow.
36 Barber, Journeys of Discovery (2005) | 37 Fullan, Hill, Crevola, Breakthrough (2006)
27
development. Five percent of the district’s budget went
to professional development, and 80 percent of that
went to teachers... The only way to improve outcomes
is to improve instruction.”38 It is not just improving
systems that recognise the primacy of this challenge:
the top-performing systems do so too. Singapore
used its National Institute of Education to deliver
high-quality professional development to its teaching
workforce: “You can have the best curriculum, the
best infrastructure, and the best policies, but if you
don’t have good teachers then everything is lost...
We provide our teachers with 100 hours of professional
development each year... If you do not have inspired
teachers, how can you have inspired students?”39
In England too, reforms focused on improving
classroom practice. As one policymaker reflected,
“Between 1988 and 1998, [many things] were changed,
changed utterly, sometimes twice or three times. And
then I’d go into a primary school classroom in 1998 and
I’d think to myself – this is very like 1988... Since 1998
we have changed that. We have taken reform inside the
classroom.”40
Certain interventions for improving instruction had a
dramatic impact on student outcomes. In just six years,
Boston increased the number of its students meeting
the MCAS standard from 25 percent to 74 percent in
Math, and from 43 percent to 77 percent in English.
In England, where there had been little or no improvement in student outcomes in literacy and numeracy for
nearly half a century, the government rolled out new
national training programs which employed bestpractice training techniques. In just three years, they
increased the number of students meeting the target
standards in literacy from 63 percent to 75 percent
(Exhibit 17).
Individual teachers need to be motivated to make the
necessary improvements. In general, this requires a
deeper change in motivation that cannot be achieved
through changing material incentives. Such changes
come about when teachers have high expectations, a
shared sense of purpose, and above all, a collective
belief in their common ability to make a difference to
the education of the children they serve.
Many of the reforms we studied were unable to deliver
substantial improvements largely because they did
not get all of these three things to happen at the same
time. While certain reforms increased accountability or
introduced performance-based incentives to improve
motivation, they did so without providing teachers with
the awareness of their weaknesses or knowledge of best
practices.
Necessary but not sufficient
Top-performing systems are relentless in their focus on
improving the quality of instruction in their classrooms.
Yet this focus on instruction, though a necessary
condition, is in itself insufficient to bring about improvement. In order to improve instruction, school systems
needed to find ways to change fundamentally what happens in the classrooms. At the level of individual teachers, this implies getting three things to happen:
Individual teachers need to become aware of
specific weaknesses in their own practice. In most
cases, this not only involves building an awareness
of what they do but the mindset underlying it.
Individual teachers need to gain understanding
of specific best practices. In general, this can only be
achieved through the demonstration of such practices
in an authentic setting.
38 Interview: Boston, January, 2007 | 39 Interview: Singapore, April, 2007 | 40 Barber, Journeys of Discovery (2005) | 41 Harvey-Beavis, Performance Based Rewards for Teachers (2003). CTAC, Catalyst for Change: Pay for Performance
in Denver (2001) | 42 Elmore, School Reform From The Inside Out (2004)
There is plenty of evidence to suggest that without
all three things in place, change will be limited. For
instance, studies which evaluated the effect of performance-based pay on student outcomes in North Carolina,
Denver and Texas show that although student outcomes
might improve to a certain extent in some schools as a
result, these gains were not substantial.41 Reforms that
expose teachers to best practices through workshops
or written materials but that do so without making this
knowledge precise enough for teachers to understand
how to apply it in their own classroom also fail: “The
notion that external ideas by themselves will result in
changes in the classroom and school is deeply flawed as
a theory of action.”42 Despite the evidence, and the fact
that almost every other profession conducts most of its
training in real-life settings (doctors and nurses in hospitals, clergy in churches, lawyers in courtrooms, consultants with clients) very little teacher training takes place in
the teacher’s own classrooms, the place in which it would
be precise and relevant enough to be the most effective.
Different approaches
There are broadly four approaches high-performing
school systems use to help teachers improve instruction,
create awareness of weaknesses in their practice,
provide them with a precise knowledge of best
practice, and motivate them to make the necessary
improvements.
Building practical skills during the initial training:
Several high-performing and improving systems
have moved their initial period of training from the
lecture theatre to the classroom. This allows them
to build teaching skills more effectively. On the
one-year Teacher Residency program in Boston,
for example, trainees spend four days each week in a
school. In England, two thirds of the time on one-year
teacher training courses is devoted to teaching practice.43 In Japan, teachers spend up to two days a week
in one-on-one coaching in their classrooms, during
their first year of training.
•Placing coaches in schools to support teachers:
All top systems, including the rapidly improving ones,
recognize that if you want good teachers, you need
to have good teachers train them, and this requires
focused one-on-one coaching in the classroom.
Expert teachers are sent into the classroom to
observe and provide one-on-one coaching in terms
of feedback, modelling better instruction, and in
helping teachers to reflect upon their own practice.
In England, teachers with a track record of excellent
instruction are given reduced teaching loads in order
to allow them to spend more time coaching their
colleagues. In Chicago and Boston, literacy coaches
work one-on-one with teachers in classrooms to help
them to improve their instruction.
despite the evidence, and the fact that
almost every other profession conducts most of
its training in real-life settings (doctors
and nurses in hospitals, clergy in churches,
lawyers in courtrooms, consultants with clients)
very little teacher training takes place in the
teacher’s own classrooms, the place in which it
would be precise and relevant enough to be the
most effective.
Selecting and developing effective instructional
leaders: Coaching is effective as an intervention,
but it can become even more so once schools have
developed the culture of coaching and development that will sustain it. To achieve this goal, certain
school systems have ensured that their school leaders are also ‘instructional leaders’. They have put in
place mechanisms for selecting the best teachers to
become principals, and then train them to become
instructional leaders who then spend a good portion
of their time coaching and mentoring their teachers.
Principals in small schools in most of the top systems
spent 80 percent of the school day focused on improving instruction and demonstrating a set of behaviours
which build the capacity and motivation of their teachers to constantly improve their own instruction.
•Enabling teachers to learn from each other:
Finally, some of the best systems have found ways
to enable teachers to learn from each other.
Teachers in most schools work alone. In a number of
the top systems, particularly those in Japan and
Finland teachers work together, plan their lessons
jointly, observe each others’ lessons, and help each
other improve. These systems create a culture in their
schools in which collaborative planning, reflection
on instruction, and peer coaching are the norm and
constant features of school life. This enables teachers
to develop continuously.
Most of the top systems combine two or three of these
approaches. While the first two approaches are interventions that improve instruction but which do not attempt to
embed a culture of continuous improvement, the other
two complement them by focusing on the creation of a
culture that can help ensure sustained improvement.
Building practical skills
during initial training
Teachers develop the bulk of their instructional capability during their first years of training and practice. In
several of the school systems we studied, the evidence
suggests that the support given to teachers during this
period (both in their initial training, and the support
they were given during their first years of practice) was
rarely as effective as it should have been. Research
shows that in the United States many teacher education
programs have little impact on teacher effectiveness.44
Frequently, this is because the connection between what
the trainee teachers do during their training, and what
they are expected to be able to do once they arrive in
the classroom, is not strong enough. Angus McBeath,
former superintendent of Edmonton’s schools in Alberta,
noted, “We would never turn out a freshly minted doctor
and say, ‘go operate on somebody’ without three or four
years of practice - guided practice.But we turn out
teachers, put them in classrooms, and ignore them.”45
43 They stipulate 18 weeks for primary postgraduate certificate programs, 24 weeks for secondary and key stage
2/3 programs.44 Chaney, Student outcomes and the professional preparation of 8th grade teachers, Goldhaber and
Brewer, Does certification matter? | 45 McBeath, Getting Districtwide Results (2006)
All of the better school systems we studied had
integrated practicum into their teacher training
programs. Boston, England, Finland and Japan went
further, in increasing the amount of intensive practical
support given to new teachers and in finding ways to
ensure that the support they give is more effective.
29
Boston: Boston has introduced a graduate teacher
training program based on a medical-residency
model, combining a large amount of practical
experience, a strong theoretical background, and
a higher-level (masters) degree qualification. After
an initial six-week summer school, trainee teachers spend one year on an apprenticeship in schools.
During this year they spend four days each week
working with an experienced teacher, and one day a
week doing coursework. During their second year,
each new teacher is allocated a mentor who provides
two-and-a-half hours of in-class coaching each week.
Mentors “model, co-teach, observe and help with
classroom management, lesson planning and instructional strategies.”46 In order to improve the quality of
mentoring on the program, Boston now employs a
number of full-time specialist mentors, each of whom
supports 14 new teachers.
England: England has placed all funding for teacher
training under the control of a new agency, the Training and Development Agency for Schools (TDA). The
TDA set strict standards for teacher training institution,
including a minimum requirement of 24 weeks47 of
practical experience on most courses (two thirds of
the total course time on one year programs) with the
requirement that this classroom experience provides
a good learning environment for trainee teachers.
Providers are inspected by an independent inspectorate; the TDA reduces funding or closes down
providers which do not meet the standards. England
has also introduced an induction year, during which
new teachers are given increased support and supervision, a reduced teaching load that allows extra time
for planning and training, and a regular performance
review to highlight areas requiring improvement.
teaching load so that they can support and coach other
teachers. Singapore appoints senior teachers and master
teachers to lead the coaching and development of the
teachers in each of its school.
Finland: Most faculties of education manage their
own training schools: these are fully operational
schools where students carry out their initial teaching
practice. The organizational structure helps to ensure
that the content of teacher training is tightly linked
to the actual practice within schools, and provides
additional opportunities for the faculty to incorporate
observation and practice gained in the classroom into
their teacher training courses.
Coaching interventions can lead to a substantial
improvement in outcomes in a short time. Through its
National Literacy and Numeracy Strategies, England has
trained numeracy and literacy coaches in every primary
school. It developed a network of national experts to train
these coaches, focusing both on effective pedagogies to
be used to improve student outcomes and on the
techniques to get teachers to employ them. The result
has been a significant improvement in outcomes over a
period of just three years. Several of the Middle Eastern
systems have used coaching strategies to effect significant changes in instruction in their schools, bringing in
coaches from foreign school systems to quickly train
large numbers of teachers in different teaching styles.
Japan: The teacher preparation programs at Japan’s
universities focus mainly on building the intrinsic
capabilities, content knowledge, and the pedagogical knowledge of aspirant teachers. In 1989, Japan
introduced an intense training program for first-year
teachers during which trainees develop their practical
teaching skills. In this program, trainee teachers work
full-time in schools and during their first year are provided with up to two days of one-on-one coaching and
support each week from ‘guidance teachers’. Guidance teachers coach and mentor but do not evaluate
new teachers during their first year in the classroom.
Placing coaches in schools to
support teachers in the classroom
The next challenge is to make in-service training an effective tool to improve instruction. Several of the systems
do this through on-the-job coaching. Expert teachers,
trained in how to coach other teachers, enter classrooms
to observe teachers, give feedback, model instruction,
and share in planning. In some cases the experts are
full-time coaches employed by the district or ministry, in
others they are experienced teachers with a track record
of excellent instruction who have been given a reduced
Selecting and developing
effective instructional leaders
The research on school leadership suggests that “school
leadership is second only to classroom teaching as an
influence on learning.”48 Some 97 percent of schools in
England rated good or excellent overall by the independent inspectorate are led by management teams that
are also rated good or excellent overall; only 8 percent
of schools with leadership teams rated satisfactory or
below are rated good or excellent overall.49 Research
shows that without an effective headteacher [principal],
a school is unlikely to have a culture of high expectations, or strive for continuous improvement. “...Schools
are vulnerable where a formerly good headteacher
becomes less effective over time, or where a strong
headteacher leaves the school without having developed
a confident and effective leadership team.”50
46 Interview: Boston, January, 2007 | 47 See note 36 | 48 NCSL, Seven Strong Claims about Successful School
Leadership (2006) | 49 Ofsted, School Inspection Data (2005/2006) | 50 National Audit Office, Improving Failing
Schools (2006)
Getting the right teachers
to become principals
To produce effective school leaders, school systems
first need to select the right people to become leaders.
Research on effective school leadership shows that,
“a small handful of personal traits explain a high
proportion of the variation in leadership effectivness.”51
To get the right people to become school leaders,
high-performing school systems provide the right
incentives to get the best teachers to apply for
leadership positions, and implement processes effective in selecting the best of those who apply. How they
do this depends mainly on whether principal selection is
centralized (i.e. controlled by the district or ministry), or
decentralized (i.e. controlled by individual schools). Singapore and Chicago illustrate two systems for doing this.
The evidence suggests that strong school leadership
is particularly important in producing improvement.
Reforms in Boston, England, and Singapore all
demonstrate that good leadership in schools is
important in effecting fast and substantial changes to
practice. Top-performing school systems leverage
a substantial and growing knowledge about what
constitutes effective school leadership to develop their
principals into drivers of improvement in instruction.
In general, developing effective instruction leaders in
schools meant doing three things:
Getting the right teachers to become principals
Developing instructional leadership skills
Focusing each principal’s time on instructional
leadership
51 NCSL, Seven Strong Claims about Successful School Leadership (2006) | 52 NCSL, Seven Strong Claims about
Successful School Leadership (2006) | 53 DfES, Independent Study into School Leadership (2007)
S
ingapore: Principals salaries are high, partly in
recognition of the demands of the role, as well as to attract strong candidates. As part of the stringent selection process for principals, candidates are put through
an Assessment Centre, which is a series of carefully
designed exercises that elicit observable behaviours
related to the core competencies of a school leader.
Candidates that are found to have principalship potential attend a six-month program run by the National
Institute of Education. These candidates are assessed
continuously by the training team, and this assessment is fed into the selection process. This ongoing
assessment over a six-month period provides a more
accurate reading of the intrinsic capabilities than is
achieved by a regular recruitment process. At the
end of the six-month program, only candidates who
are found to be ready for principalship and can be
matched to schools are appointed as principals.
Chicago: Principals are selected and employed
by individual school committees, making it more
difficult for the district to control quality than in
Singapore. In response to this organizational
challenge, the city has introduced tough eligibility
criteria, creating a two-stage selection process.
In order to apply for a principal position, candidates
first need to pass through this eligibility process (twothirds of applicants fail on their first attempt). Eligible
candidates then compete for principal
positions at individual schools (Exhibit 18).
Developing instructional
leadership skills
Getting the right people to become school leaders
is very important, but so is providing these people
with the right set of skills to be effective leaders.
Essentially, all successful school leaders “draw on the
same repertoire of basic leadership practices.”52 The
best-performing school system’s implement a coherent and aligned development model, (frequently based
on an apprenticeship model) which helps aspiring and
existing school leaders to develop these practices
(Exhibits 19 & 20).
Focusing each principal’s time
on instructional leadership
Once the school system has identified and developed
the right people with the right skills, it then needs to
structure the roles, expectations and incentives to ensure that its principals focus on instructional leadership,
not on school administration. This contrasts with school
systems in which many principals spend most of their
time on tasks not directly related to improving instruction
in their schools, thus limiting their capacity to effect real
improvement in student outcomes.53 The systems which
seek to use their principals as drivers of reform expect
31
them to be excellent instructors who spent most of their
time coaching teachers. In the words of one highly
successful principal we interviewed: “Being a teacher
is about helping children to learn. Being a principal is
about helping adults to learn. That’s why it’s tough...
I walk the halls, walk the halls, and walk the halls... I only
look at my inbox after everybody else leaves.”54
Enabling teachers to learn
from each other
The final approach is to enable teachers to learn from
each other. Unlike other professions, where professionals naturally operate in teams, teachers generally work
alone, denying them natural opportunities to learn from
each other. Several school systems employ strategies
aimed to change this by creating schools in which teachers regularly observe each others’ practice, thereby
producing an environment which stimulates the sharing
of knowledge on what works and what does not, encourages teachers to give each other feedback, and helps
shape a common aspiration and motivation for improving the quality of instruction. These systems are some of
the best performing of all of the systems we studied.
Japan: The learning culture in its schools is centred
on ‘lesson study’ (kenkyuu jugyou). Groups of teachers work together to refine individual lessons, jointly
planning, executing and then evaluating different
instructional strategies for achieving a specific
learning objective. Groups of teachers visit each
others classrooms to observe and understand the
practice of other teachers (Exhibit 21). There is a
strong emphasis on making sure that best practices
are shared throughout the school: “When a brilliant
American teacher retires, almost all of the lesson
plans and practices that she has developed also
retire. When a Japanese teacher retires, she leaves
a legacy.”55
Boston: Teachers are timetabled so that all of the
teachers who teach the same subject at the same
grade level have ‘free classes’ together. This time
is used for jointly planning and analysing teaching
practice based on assessment data. The sessions are
facilitated, either by the principal or one of the literacy
coaches, and use assessment data as the basis for
structured discussion. The aim is to uncover
differences between the instructional practices of the
various teachers in the school and to understand how
these differences impact results. The sessions are
followed by peer observation and common
planning of teaching strategies (Exhibit 22). Some
of the schools using this approach are built on an
open plan: without doors between classrooms, and
sometimes without walls. This facilitates collaborative
teaching and encourages teachers to learn from each
other.
54 Interview: Boston, January, 2007 | 55 Chenoweth, from Lewis, Does Lesson Study have a future in the
United States (2002)
Conclusion
Many of the reforms we studied failed to deliver
improvement because they had little effect on what
happened inside the classroom. Cuban’s analogy of the
effect of many school reforms on teaching practice is that
they have a similar effect to that of a storm on the ocean:
“The surface is agitated and turbulent, while the ocean
floor is calm and serene (if a bit murky). Policy churns
dramatically, creating the appearance of major changes
.. while deep below the surface, life goes on largely uninterrupted.”56
Finland: Teachers are given one afternoon each
week for joint planning and curriculum development.
The fact that the national curriculum specifies only
general outcome goals, rather than the path by which
to attain them, mean that teachers in schools have
to work together to develop the curriculum and the
instructional strategies tailored to the needs of their
school. Schools in the same municipality are encouraged to work together and share materials so that
best practices spread quickly throughout the system.
All the evidence from both the high- and low-performing systems shows that the most effective way to deliver
sustained and substantial improvements in outcomes
is through sustained and substantial improvements in
instruction. School systems from Singapore to England
and from Finland to Boston have done this successfully,
catalysing significant improvements in instruction that
have led to demonstrable improvements in student outcomes. The four different approaches that have proved
effective all begin with an understanding of what it takes
to improve the quality of instruction of a single teacher,
and then develop the systems to create these conditions
for all teachers. They show that while the task of transforming instruction on a large scale is challenging, it is
nevertheless achievable.
being a teacher is about helping children to
learn. Being a principal is about helping
adults to learn. that’s why it’s tough...
I walk the halls, walk the halls, and walk
the halls... i only look at my inbox after
everybody else leaves
33
“Delivering for
every child”
Getting the right people to become teachers and
developing them into effective instructors gives school
systems the capacity they need to deliver the improved
instruction that leads to improved outcomes.
High-performing school systems go further than this
and put in place processes which are designed to
ensure that every child is able to benefit from this
increased capacity. These systems set high expectations
for what each and every child should achieve, and then
monitor performance against the expectations,
intervening whenever they are not met. High-performing school systems construct effective interventions at
the level of the school, identifying schools that are not
performing satisfactorily, and intervening to raise
standards of performance. The very best systems
intervene at the level of the individual student,
developing processes and structures within schools
that are able to identify whenever a student is starting
to fall behind, and then intervening to improve that
child’s performance.
The extent to which a school system is able to realise the
benefits of improved instruction depends on its ability
to deploy it effectively: the system needs to ensure that
every child, rather than just some children, has access to
excellent instruction. Ensuring that every child benefits
from high-quality instruction is not only an important
end in itself, the evidence from international assessments
suggests that strong performance for the system as a
whole is dependent on this being the case. For example,
the PISA scores of the top performing systems show a
low correlation between outcomes and the home background of the individual student (Exhibit 23). The best
systems have produced approaches to ensure that the
school can compensate for the disadvantages resulting
from the student’s home environment.
for example, the PISA
scores of the top
performing systems
show a low
correlation between
outcomes and the home
background of the
individual student
57 Michael Barber, lecture, London, 2007
58 RAND, The Economics of Investing in Universal ­Preschool Education in California (2005)
In many of the systems we studied, the systems to
ensure consistent high-quality instruction are either absent or broken. In England, for instance, mechanisms to
intervene in poorly performing schools were introduced
into the system only relatively recently: “The idea that a
school could be failing, known to be failing, and left
failing, looks scandalous in retrospect... High
performance requires every child to succeed”57.
Systems that compensate for the effects of low income
and poor home background on educational achievement are still far from universal. Yet in many cases these
interventions are critical in ensuring that the overall level
of performance of the school system can be raised
sufficiently: the evidence suggests that poorer families
and poorer localities invest less in their children’s
education, dragging down overall performance. In the
United States, for instance, children whose mothers have
bachelor’s degrees are almost twice as likely to be
enrolled in preschool programs as children whose
mothers have not completed high school.58 In general,
richer localities produce better schools because of
better funding. Annual current expenditure on students
in the top five percent, best-funded public schools in the
35
United States, for instance, is $ 12,400, while expenditure
in the bottom five percent, worst-funded schools it totals
just $ 5,700. These discrepancies also impact recruitment. Students in schools that serve students from a
poorer socioeconomic background are twice as likely
to be taught by teachers with less than three years
experience than students in schools which serve
students from a richer socioeconomic background.59
All these factors compound the lower expectations and
inequality of opportunity for students from a poorer
background.
The high-performing systems are better at ensuring that
each student receives the instruction they need to
compensate for their home background. They start by
setting clear and high expectations for what individual
students should know, understand, and be able to do.
They ensure that resources and funding are targeted at
those students who need them most, not those who need
them least. They then closely monitor the performance
of schools against these expectations and develop
effective mechanisms for intervening when these
expectations are not met. Different systems have
different ways of doing this. In general, the level of monitoring and intervention in the best-performing systems is
inversely proportional to the capacity of individual
teachers and the schools to improve by themselves.
The best systems locate the processes for monitoring and
intervention in the schools themselves, where they are
able to identify the students in need of support and provide that support when needed on a continuous basis.
Setting high expectations for
what students should achieve
All of the top-performing and rapidly improving systems
have curriculum standards which set clear and high
expectations for what students should achieve. Boston’s
reform is directed towards increasing the number of students meeting the Massachusetts state standards – some
all of the top-performing and rapidly
improving systems have curriculum standards
which set clear and high expectations for what
students should achieve
of the toughest in the United States. Alberta sets high
expectations for student achievement, and then
participates in international exams such as PISA and
TIMSS to benchmark its standards: “If our kids are
passing the provincial examinations but performing
below other top systems in PISA, then we know that
we need to raise our standards.”60 Finland reformed its
curriculum in 1992, replacing a previously rigid national
curriculum with targets for all students: “We do well
because we aim high.”61
In general, school systems use more prescriptive
standards when the overall performance level of the
system is low, and then relax those standards as the
system improves. For instance, Boston found that the
Massachusetts state standards were too loose, given
the overall performance of its system: “The standards
were at 10,000 feet, we needed something closer to the
ground.”62 England’s current National Curriculum is twothirds the length of when it was originally introduced
in 1990, its length reflecting a growing willingness to
give teachers more freedom as quality of outcomes has
improved. Finland, one of the highest performing of
the systems we studied, arguably has one of the least
prescriptive curricula: “The target is high, but we want
teachers to be able to make their own choices.”63 The
Finnish curriculum emphasizes the need for teachers to
adapt learning to the specific context in which they find
themselves, and recognizes the fact that children learn
at different rates, while at the same time setting high
expectations for what should ultimately be achieved.
The process by which these expectations are set is
frequently long, difficult and controversial, and the
resulting curricula varies widely as a result. Yet some
things are a constant in all the top systems. They all
place a strong focus on numeracy and literacy in the
early years based, in part, on substantial research
evidence which shows that early ability in core skills
is strongly correlated with a range of future outcomes:
a major longitudinal study in the United Kingdom, for
instance, found that test scores in literacy and numeracy
at age seven were significant determinants of earnings at age 37, even after controlling for socioeconomic
background.64 There is also a growing tendency to align
standards globally, particularly in reference to those of
the OECD’s PISA assessments and other leading school
assessment systems. Certain systems try to match
current teaching to the country’s future requirements.
Singapore has invested heavily in trying to anticipate
the required range and mix of skills that its students will
need when they graduate to further grow Singapore’s
economy, and matches its curriculum to those needs.
Whatever the differences, however, all the top systems
recognise the need to set clear and high expectations
for the performance of their students.
Monitoring and intervening
at the schools level
All of the top-performing systems also recognize that
they can not improve what they do not measure.
Monitoring outcomes allows them to identify and spread
59 Kati Haycock, Achievement in America: Can we close the gaps (2006) | 60 Phone interview: Alberta, March, 2007 | 61 Interview: Finland, March, 2007 | 62 Interview: Boston, January, 2007
63 Interview: Finland, March, 2007 | 64 Currie, Thomas, Early Test Scores, Socioeconomic Status and Future Outcomes (1998)
all of the
top-performing
systems also
recognize that
they can not
improve what
they do not
measure
65 Phone interview: May, 2006 | 66 Interview: New Zealand, May, 2006
best practices, to pinpoint areas of weakness, and to
hold schools accountable for their results. In general, the
intensity of the monitoring that is carried out is in inverse
proportion to the overall performance, both within
and between systems. Thus, while rapidly improving
systems such as Boston and Chicago test every student
every year between grade three and grade eight,
top-performing systems such Finland have largely
dispensed with national examinations, conducting only
periodic assessments of student performance, the results of which stay confidential. Within systems, schools
which perform well are subject to less monitoring (for
instance, Singapore exempts its top schools from certain
examinations), whereas schools which perform poorly
are subject to more intensive scrutiny (for instance,
schools in England which are identified as underperforming are subject to more frequent reviews until their
performance improves).
The high-performing systems use two mechanisms
for monitoring the quality of teaching and learning
(Exhibit 24):
Examinations: Examinations test what students
know, understand and can do, providing an objective
measure of actual outcomes at a high level of detail.
Examinations also have a powerful effect in driving
the performance of any school system. In the words
of one Australian educationalist, “What gets tested is
what gets learnt, and how it is tested determines how
it is learnt.”65
School review: School reviews, or inspections, assess
the performance of a school against a benchmark
set of indicators. Unlike examinations, they measure
both outcomes and the processes which drive them,
and as a result, can help schools and systems identify
specific areas which are in need of improvement.
School reviews also enable systems to measure some
of the more subtle and complex desired outcomes of
a school system, which are difficult or impossible to
measure in examinations.
In many of the top-performing systems, responsibility
for monitoring outcomes has been separated out from
the responsibility for improving those outcomes. In the
words of one policymaker in New Zealand, “You can’t
have the same people who are responsible for improving education be responsible for judging whether or
not that improvement has occurred.”66 Hong Kong has
created a school inspectorate, which is separate to the
school branch offices to which the schools report, but still
inside its ministry of education, and an independent
examination board (HKEAA), which is outside the
ministry but is still ultimately accountable to the minister.
England has created an independent inspectorate
(Ofsted) which is directly accountable to parliament, and
it places national assessments under a semi-independent regulator (the QCA). New Zealand has created an
independent schools inspectorate which reports to its
own minister (though the two portfolios – Minister of
Education and Minister Responsible for the ERO – are
frequently held by the same person).
In general, the arrangements for school review depend
on the overall level of performance of the system, and in
some cases, the individual performance level of schools.
Typically, as the school system improves, the task of
monitoring moves from external agencies to the schools
themselves.
Annual external review: School systems
embarking on ambitious reforms tend to use more
frequent external reviews. In New York, Qatar, and
Bahrain (all of which are embarking on ambitious
reform efforts) all the schools are to be reviewed by
an external inspectorate once every year. All three
systems plan to reduce either the length or the frequency of external reviews as their system improves.
37
Self-evaluation with external review every 3-4
years: In England, Hong Kong and New Zealand,
schools are inspected once every three-to-four years,
with a strong emphasis on on-going school self-evaluation during the intervening period. All three systems
are evolving towards less intensive review models
as they improve: England, for instance, introduced a
new inspection regime in 2005 which in most cases
more than halved the number of days’ spent on the
inspection. Schools which perform well are inspected
less frequently and less intensively than those which
perform badly.
S
elf-review with occasional external review:
In Singapore, schools are expected to undertake
regular self-evaluation: external school reviews
occur only once every five years. In Finland, there
is no formal review cycle: schools can request an
informal audit of their teaching and learning at any
point to complement their own internal review
processes.
Monitoring outcomes ensures that the system has the
information it needs to be able to intervene when schools
start to fail. Effective interventions, best illustrated by
those conducted in England, New York and New
Zealand, are characterised by a number of features:
Publication of performance reports: In many cases,
systems that set out to be transparent about the
performance of their schools (typically by publishing
the inspection or examination data) create greater
public accountability and awareness which, in turn,
drives further improvement. In the words of one New
Zealand policymaker: “[We] make everything public; it creates tension in the system – transparency
over the problems – and that drives improvement.”67
However, the evidence from the systems which
publish performance reports shows that though many
good schools improve further under the pressure
you can’t have the
same people who are
responsible for
improving education
be responsible for
judging whether or
not that improvement
has occurred
resulting from the transparency of the system, failing
schools seldom improve for this reason alone. “If a
school does not know how to improve, if it lacks the
capacity to improve, then no amount of pressure will
change instruction.”68 Indeed, in some top-performing systems, transparency about school performance is perceived as an obstacle rather than an aid
to improvement: “Improvement comes from building
capacity, and harnessing the motivation that teachers
and schools already have; additional pressure just
leads to regressive behaviours [for example teaching
to the test, drilling students on examination questions,
preventing poor students from taking the test, and
potentially fraudulent behaviour].”69 Finland keeps
performance assessments and audits confidential,
providing results only to the school that has been
assessed and to their municipalities. Hong Kong has
adopted a policy of not publishing performance data
in order to reduce what is widely perceived to be
the already excessive performance pressures on
students and teachers.
Funding: New Zealand, Alberta, England and
Chicago have all introduced funding models which
divert additional resources to those schools which are
in need of improvement. The funding formulae
provides increased funding to schools which
enrol pupils from disadvantaged backgrounds.
England has made additional funds available to
a large number of schools perceived to be at a
higher- than-average risk of failure: this sum totals
$1.5 billion dollars each school year.
Intervention to replace or improve leadership:
Most evidence about remedying the performance of
failing schools suggests that strong school leadership
is essential. The top-performing systems, as well as
rapidly improving ones, create mechanisms to allow
central or local government to replace the school’s
leadership in cases where normal governance
arrangements do not allow this to happen. In Chicago,
England, and New Zealand, the district, local authorities, or central government, respectively, have the
right to replace the school leadership when a school
fails to improve. Boston removes the bottom five
percent of principals during the first year of their
reform, and then several of the lowest performing
principals each year thereafter.
In addition, the best systems use the results of monitoring and intervention to identify best practices, which can
then be spread throughout the system. Singapore
studies the practices in its best schools, and has ensured
that the lessons from this are transferred to other schools.
Singaporean researchers have built classroom-laboratories at the National Institute for Education where they
carefully monitor student reactions to new instructional
ideas, techniques and strategies being tested there.
They then apply their findings to future education
reform. Singapore spends almost US$10 million each
year on research into better instructional practice.
England uses data from its inspections and assessments
to identify the best schools and teachers, and then uses
this to develop new approaches and further reform.
67 Ibid | 68 Michael Barber, op.cit | 69 Interview: Finland, March, 2007
Finnish children start preschool at age six and school at
age seven, three years later than many of their European
counterparts. Once in primary school, they study for just
four to five hours a day. Finnish children receive fewer
hours of instruction between the ages of seven and
14 than any other children in an OECD country. Yet by
age 15, Finnish children top the world in the OECD’s
assessments of reading, mathematics, science and
problem solving, performing significantly better than
all their Scandinavian neighbours.
Monitoring and intervening
at the students level
Intervention at the level of the school prevents clusters of
failure from emerging in the system. However, the most
effective schools and school systems monitor and intervene at the level of individual students. This is essential
if the system is to deliver consistently strong performance throughout all its schools. Evidence from the United
States shows that by age three the average child of
professional parents has a vocabulary of 1,100 words
and an IQ of 117, whereas the average child of parents on
welfare has a vocabulary of just 525 words and an IQ of
79.70 Unless schools intervene effectively to compensate
for the impact of a poorer home environment, they stand
little chance of closing this gap. The best schools in each
system have developed mechanisms for doing just this.
Finland has gone further than any other system in
ensuring that there is a uniformly high performance
across its entire system.
70 Hart and Risley, from New York Times “What it takes to Make a Student” (26 November 2006).
Part of the explanation for this is that Finland gets the
right people to become teachers (recruiting from the top
10 percent of school leavers, controlling admission to
teacher education, and paying good starting compensation), and has developed them into effective instructors
once they are selected (through excellent pre-service
training, excellent instructional leadership, and professional learning communities within schools). This is not
the entire story, however. Finland has also developed
a highly effective system of interventions to support
individual students within schools. Each Finnish school
employs a number of special education teachers.
In the schools we visited during our benchmarking,
we observed that on average there was one special
education teacher for every seven class teachers.
Special education teachers provide support one-on-one
or in small-group to students who are at risk of falling
behind. They intervene to support 30 percent of all
students in a school in any given year. These special
education teachers provide support mainly in the
subjects of Finnish and mathematics, and are given an
additional year of teacher training to support them in
this role.
Special education has been de-stigmatized in Finland
by two practices. Firstly, by the high volume of students
who take part in the program. Secondly, by the practice
in which the best students are also sent, on occasion, for
additional instruction: this makes it clear that such intervention is not necessarily a sign of underperformance.
By intervening quickly at the level of individual students,
Finland prevents early failure compounding into
long-term failure, and thus has found a way to maintain
strong and consistently equitable outcomes in its schools
(Exhibit 25).
Other top-performing systems have developed different
approaches by which they ensure that they can intervene to support children who are falling behind. Asia’s
systems depend on strong commitment from individual
teachers to provide the necessary extra support where it
is required. In Singapore, for instance, teachers typically
remain in school for several hours after formal lessons
have ended, providing additional teaching to those students who need it most. Singapore also provides extra
classes for small groups of the lowest-performing 20
percent of students during the first and second grades.
In New Zealand, the Reading Recovery program is
designed to provide extra instruction for students whose
reading performance is poor.
Conclusion
A combination of monitoring and effective intervention
is essential in ensuring that good instruction is delivered
consistently across the system. High-performing
school systems monitor their performance through
examinations and inspections, making the intensity of
this monitoring inversely proportional to the capacity
of individual schools to improve by themselves.
They use the results of the monitoring to inform effective
interventions to raise standards and achieve a uniformly
high performance. The best systems take these
processes inside schools, constantly evaluating student
performance and constructing interventions to assist
individual students in order to prevent them from falling
behind.
39
Conclusion:
The system and
the journey
South Korea and Singapore demonstrate that a school system can go from
low performance to high performance within a few decades. This achievement is even more remarkable given that it typically takes a long time to
see the impact of a reform effort (the test scores on graduation from high
school are highly dependent on the quality of primary education that
students received ten years earlier which, in turn, is highly dependent on
the quality of people who became teachers sometime before this). Boston
and England have also demonstrated that substantial improvements in
both the outcomes and the factors that drive them (for instance, the status
of the teaching profession) can be achieved in short period of time.
All the different school systems that have improved significantly have
done so primarily because they have produced a system that is more
effective in doing three things: getting more talented people to become
teachers, developing these teachers into better instructors, and in
ensuring that these instructors deliver consistently for every child in the
system. The way in which they have done these things varies somewhat.
Singapore’s school system is managed from the centre and they have
used this to drive through improvements in performance. In England,
policymakers have relatively less control over its more decentralized
school system, so they have used standards, funding, public accountability, and strong support mechanisms to create the conditions under which
improvement can occur. In other systems, the strength of unions or other
political actors has had influence over the pace and path of reform, though
maybe not its ultimate direction.
Putting these three things in place often requires more general reform
to the school system. School reforms rarely succeed without effective
leadership, both at the level of the system, and at the level of individual
schools. One study noted that, “there is not a single documented case of a
school successfully turning around its pupil achievement trajectory in the
absence of talented leadership.”71 Similarly, we did not find a single school
system which had been turned around that did not possess sustained,
committed and talented leadership. Changing the governance or
management of a system might, therefore, be a necessary prerequisite
for improvement, even if such changes do not necessarily lead to
improvement in themselves. Similarly, systems which do not fund
equitably ensure that poorer schools have little chance of performing well,
even though simply changing the funding structure does not of itself
necessarily lead to improvement. The nature of the curriculum is
critical, though without an effective system for delivering the curriculum,
any changes to course content or learning objectives will have little impact
on outcomes.
The school systems we have benchmarked demonstrate that delivering
substantial improvements in outcomes is both challenging and achievable. The three themes identified in this paper, and the best practices for
achieving them, form the core of what system leaders must do to ensure
improvement. The paths which the various school systems have taken in
the past, and the paths which other school systems will have to take in the
future to achieve similar performance are, inevitably, very different.
Yet all school systems need to be able to answer a similar set of questions
regarding these three themes and be able to match the existing
parameters of best performance (Exhibit 26).
In many cases, extraneous factors hold back change and these problems
need to be tackled first to enable the school system to implement policies
and processes that will improve student performance. Context, culture,
politics and governance will determine the course which system leaders
must follow, as will their point of departure. Yet, ultimately, for achieving
real improvement in outcomes, none of these things will be as important
to the school system and its leaders as three guiding principles:
1) the quality of an education system cannot exceed the quality of
its teachers, 2) the only way to improve outcomes is to improve
instruction and, 3) achieving universally high outcomes is only
possible by putting in place mechanisms to ensure that schools
deliver high-quality instruction to every child.
71 NCSL, Seven Strong Claims about Successful School Leadership (2006)
41
43
Bibliography
Allington, Johnston, What do we know about effective fourth grade teachers and their
classrooms (2000).
Aspen Institute, Annenberg Challenge, A Case Study to Support Leadership Transition
in the Boston Public Schools (2006)
Balanskat, Blamire, Kefala, The ICT Impact Report (2006).
Barber, Instruction to Deliver (2007).
Eurydice, Keeping Teaching Attractive for the 21st Century (2005).
Eurydice, Key Data on Education in Europe (2005).
Barber, Journeys of Discovery (2005).
Fuchs, Wößmann, What accounts for international differences in student performance
(2004).
Barber, Mourshed, Whelan, Improving Education in the Gulf (2007).
Fullan, The New Meaning of Educational Change (2007).
Broad Foundation, The 2006 Broad Prize for Urban Education (2006).
Fullan, Hill, Crevola, Breakthrough (2006).
Business Week, Bill Gates Gets Schooled, (2006).
Goh, Gopinathan, The Development of Education in Singapore Since 1965 (2007).
Cheng, New Paradigm for Re-engineering Education (2004).
Goldhaber and Brewer, Does certification matter?
Cross City Campaign, A Delicate Balance: District policies and Classroom Practice
(2005).
Guarino, Santibanez, Daley, Teacher Recruitment and Retention (2006).
CTAC, Catalyst for Change: Pay for Performance in Denver (2001).
Cuban, How teachers taught: Constancy and change in American classrooms, 18901980 (1984).
Currie, Thomas, Early Test Scores, Socioeconomic Status and Future Outcomes (1998).
Decker, Mayer, Glazerman, The Effects of Teach For America: Findings from a National
Evaluation (2004).
DfES, Independent Study into School Leadership (2007).
Ellis, The Training and Development Agency for Schools (2006).
Elmore, School Reform From The Inside Out (2004).
Elmore, Grossman, King, Managing the Chicago Public Schools (2006).
Eurydice, Evaluation of Schools Providing Compulsory Education in Europe (2004).
Hanushek, The Evidence on Class Size (2003).
Hanushek, Wößmann, The Role of Education Quality in Economic Growth (2007).
Hart and Risley, from New York Times “What it takes to Make a Student” (26 November
2006).
Harvey-Beavis, Performance Based Rewards for Teachers (2003).
Institute of Education, Effective Provision of Preschool Education Project (2003).
Institute of Education Sciences, Effectiveness of Reading and Mathematics Software
Products (2007).
IPSE, An evaluation of innovative approaches to teacher training on the Teach First
Programme (2006).
Johnes and Johnes (eds.), International Handbook on the Economics of Education
(2004).
Kati Haycock, Achievement in America: Can we close the gaps (2006).
Lee, Goh, Making Teacher Education Responsive and Relevant (2007).
Leschly, The Gates Foundation and Small High Schools (2003).
Leschly, Transformation of Seattle Public Schools (2002).
Lewis, Does Lesson Study have a future in the United States (2002).
Marzano¸ A New Era of Education Reform (2000).
45
McBeath, Getting Districtwide Results (2006).
Pritchett, Educational Quality and Costs: A Big Puzzle and Five Possible Pieces (2004).
Michaelowa, Wetchler, The Cost-Effectiveness of Inputs in Primary Education (2006).
Ramsey, Quality Matters (2000).
National Audit Office, Improving Failing Schools (2006).
RAND, The Economics of Investing in Universal Preschool Education in California
(2005).
National Centre for Education Statistics, A Closer Look at Charter Schools using Hierarchical Linear Modeling (2006).
National Centre for Education Statistics, America’s Charter Schools: Results from the
NAEP Pilot Study (2003).
National Centre for Education Statistics, NEAP 2004: Trends in Academic Progress
(2005).
NCEE, Tough Times or Tough Choices (2007).
NCSL, A Model of School Leadership in Challenging Urban Environments (2006).
NCSL, Seven Strong Claims about Successful School Leadership (2006).
NCTQ, Increasing the Odds: How good policies can yield better teachers.
New York City Department of Education, Children First (2007).
NFER, Trends in Standards in Literacy and Numeracy in the United Kingdom (1997).
OECD, Attracting, Developing and Retaining Effective Teachers (2005).
OECD, Education at a Glance (2000-2006).
OECD, PISA: Learning for Tomorrow’s World (2004).
Ofsted, School Inspection Data (2005/2006).
Phillips, Linking NAEP Achievement Levels to TIMSS (2007).
Rivkin, Hanushek, Kain, Teachers Schools and Academic Achievement (2005).
Rothman, Achievement in Literacy and Numeracy by Australian 14-year olds (2002).
Sanders & Rivers, Cumulative and Residual Effects of Teachers on Future Student Academic Achievement (1996).
Schleicher, Why Education is Key for Europe’s Success (2006).
Scientific American, Does Class Size Matter (2001).
Swanson, A Statistical Portrait of Public High School Graduation (2003).
Thompson and Wiliam, Tight But Loose (2007).
Training and Development Agency for Schools, Press Release (11 August 2005).
UNESCO, EFA Global Monitoring Report (2005).
Walker, Zhu, Education, Earnings and Productivity: Recent UK Evidence (2003).
Washington School Research Centre, The Power of Early Success (2005).
47
References
1 Spending per student in primary education, relative to GDP per capita.
15 Scientific American, Does Class Size Matter (2001).
2 In the course of the literature survey carried out as part of this research we
consulted more than 500 papers, treatises and books on teaching and school
systems. The most significant of these are referred to in the Bibliography at the end
of this report.
16 Teacher Effects on Student Achievement (1997).
3 P
ritchett, Educational Quality and Costs: A Big Puzzle and Five Possible Pieces
(2004).
17 Kati Haycock, Achievement in America: Can we close the gaps (2006).
19 Interview: South Korea, 2007.
20 NCTQ, Increasing the Odds: How good policies can yield better teachers.
4 Barber, Journeys of Discovery (2005).
21 Decker, Mayer, Glazerman, The Effects of Teach for America: Findings from a National Evaluation (2004).
5 NFER, Trends in Standards in Literacy and Numeracy in the United Kingdom (1997).
22 NCEE, Tough Choices or Tough Times (2007).
6 NAEP, America’s Charter Schools: Results from the NAEP Pilot Study (2003).
23 Interview: GCC, May 2006.
7 Business Week, Bill Gates Gets Schooled, (2006).
24 Training and Development Agency for Schools (11 August 2005).
8 Interview: New Zealand, May, 2006.
25 Allington, Johnston, What do we know about effective fourth grade teachers and
their classrooms (2000). Interviews in Singapore, South Korea, and Hong Kong.
9 C
ross City Campaign, A Delicate Balance: District policies and Classroom
Practice (2005).
10 OECD, Attracting Developing and Retaining Effective Teachers (2005).
11 Hanushek, The Evidence on Class Size (2003). Shapson, An experimental study on
the effects of class size. Akerhielm, Does class size matter?
12 Ibid.
13 The most optimistic estimates of the effectiveness of reducing class size on
student achievement suggest that a reduction in class size from 23 to 15 in the early
grades leads to an improvement in performance equivalent to 0.2 standard
deviations.
14 Sanders & Rivers, Cumulative and Residual Effects of Teachers on Future Student
Academic Achievement (1996).
26 Before 2007, the first round of the recruitment process had been based mainly on
achievement at secondary school.
27 A full salary is paid during training on one-year programs. On longer programs, a
salary is only paid during the final part of the course.
28 Interview: Finland, March, 2007.
29 OECD, Attracting Developing and Retaining Effective teachers, (from Dolton,
Wolter, Denzler) p. 70.
30 Starting primary teacher salaries in South Korea are 141 percent of GDP per capita,
rising to 389 percent of GDP per capita (compared to OECD averages of 95 percent
and 159 percent of GDP per capita respectively) (2003).
31 The increase in the maximum salary over the starting salary in high-performing
systems is as follows: an increase of 18 percent in Finland, 45 percent in New Zealand and the Netherlands, 46 percent in England, and 47 percent in Australia (average across all states and territories), compared to an OECD average of 70 percent.
Source: OECD, Education at a Glance 2005.
32 Attracting Developing and Retaining Teachers: Country Report for the
Netherlands, pp. 36-37.
33 2003 (OECD, Education at a Glance 2005).
49
34IPSE, An evaluation of innovative approaches to teacher training on the Teach First
Programme (2006).
35 Teach First targets graduates of the top universities in the United Kingdom, asking
them to spend two years teaching. It then supports them in getting other jobs in the
private sector after they had finished two years teaching. Not only are its teachers
highly successful, but 47 percent of the first cohort decided to stay in teaching after
the end of the two-year program.
36 Barber, Journeys of Discovery (2005).
37 Fullan, Hill, Crevola, Breakthrough (2006).
38 Interview: Boston, January, 2007.
39 Interview: Singapore, April, 2007.
40 Barber, Journeys of Discovery (2005).
41 Harvey-Beavis, Performance Based Rewards for Teachers (2003). CTAC, Catalyst
for Change: Pay for Performance in Denver (2001).
42 Elmore, School Reform From The Inside Out (2004).
43 They stipulate 18 weeks for primary postgraduate certificate programs, 24 weeks
for secondary and key stage 2/3 programs.
44 Chaney, Student outcomes and the professional preparation of 8th grade teachers,
Goldhaber and Brewer, Does certification matter?
49 Ofsted, School Inspection Data (2005/2006).
50 National Audit Office, Improving Failing Schools (2006).
51 NCSL, Seven Strong Claims about Successful School Leadership (2006).
52 NCSL, Seven Strong Claims about Successful School Leadership (2006).
53 DfES, Independent Study into School Leadership (2007).
54 Interview: Boston, January, 2007.
55 Chenoweth, from Lewis, Does Lesson Study have a future in the United States
(2002).
56 Cuban, How teachers taught: Constancy and change in American classrooms,
1890-1980 (1984).
57 Michael Barber, lecture, London, 2007.
58 R AND, The Economics of Investing in Universal ­Preschool Education in
California (2005).
59 Kati Haycock, Achievement in America: Can we close the gaps (2006).
60 Phone interview: Alberta, March, 2007.
61 Interview: Finland, March, 2007.
62 Interview: Boston, January, 2007.
63 Interview: Finland, March, 2007.
64 Currie, Thomas, Early Test Scores, Socioeconomic Status and Future
Outcomes (1998).
65 Phone interview: May, 2006
66 Interview: New Zealand, May, 2006.
67 Ibid.
68 Michael Barber, op.cit.
45 McBeath, Getting Districtwide Results (2006).
69 Interview: Finland, March, 2007.
46 Interview: Boston, January, 2007.
70 Hart and Risley, from New York Times “What it takes to Make a Student”
(26 November 2006).
47 See note 36.
48 NCSL, Seven Strong Claims about Successful School Leadership (2006).
71 NCSL, Seven Strong Claims about Successful School Leadership (2006).
authors
[email protected]
[email protected]